MCU board mounted in box with two serial port DB-9 connectors

[Lorem Ipsum]

I'm looking for an MCU that can be programmed for a simple translation task, that is already mounted in a small box and ready for field use.

The task is fairly simple. The main part is to receive a once per second message of around 50 characters, so low data rate, at 9,600 bps. There is a date and a time field that needs to be modified (adding colons and slashes to make them more human readable). The modified message is transmitted on the other port, also at 9,600 bps. Every 20 messages a header of 5 lines, again around 50 char per line, is sent. Still a low data rate, relative to the ~1,000 cps that can be sent.

This has been prototyped using an Arduino, with a MAX3232CPE chip on a perf board and a custom, 3D printed enclosure. But, about half the units exhibit a problem, where at some point, they will go haywire and garble the transmitted data.

This is not my project. I'm still trying to get more info on the details of what was built. In the meantime, I was looking for a box, with all the guts needed, already assembled... add code and run! But I can't seem to find anything less than a Linux or Windows box. The Arduino isn't running an OS and that seems like total overkill.


Can anyone recommend a good box level product for this? Here are some non-programmable products that would be perfect, if they supported programming of the CPU.

https://www.brainboxes.com/product/ethernet-to-serial/db9/es-257
https://www.brainboxes.com/product/usb-to-serial/usb/us-257

I've written to them to see if they can sell something like this with support for user code.

[Michael Smart]

Post by Lorem Ipsum
I'm looking for an MCU that can be programmed for a simple translation task, that is already mounted in a small box and ready for field use.
The task is fairly simple. The main part is to receive a once per second message of around 50 characters, so low data rate, at 9,600 bps. There is a date and a time field that needs to be modified (adding colons and slashes to make them more human readable). The modified message is transmitted on the other port, also at 9,600 bps. Every 20 messages a header of 5 lines, again around 50 char per line, is sent. Still a low data rate, relative to the ~1,000 cps that can be sent.
This has been prototyped using an Arduino, with a MAX3232CPE chip on a perf board and a custom, 3D printed enclosure. But, about half the units exhibit a problem, where at some point, they will go haywire and garble the transmitted data.

How about something like the STM32 'Blue Pill' ? Runs Forth e.g. Mecrisp you just need to add the MAX3232 and sounds like not a lot of Forth code.
Mike

[Lorem Ipsum]

Post by Michael Smart

Post by Lorem Ipsum
I'm looking for an MCU that can be programmed for a simple translation task, that is already mounted in a small box and ready for field use.
The task is fairly simple. The main part is to receive a once per second message of around 50 characters, so low data rate, at 9,600 bps. There is a date and a time field that needs to be modified (adding colons and slashes to make them more human readable). The modified message is transmitted on the other port, also at 9,600 bps. Every 20 messages a header of 5 lines, again around 50 char per line, is sent. Still a low data rate, relative to the ~1,000 cps that can be sent.
This has been prototyped using an Arduino, with a MAX3232CPE chip on a perf board and a custom, 3D printed enclosure. But, about half the units exhibit a problem, where at some point, they will go haywire and garble the transmitted data.

How about something like the STM32 'Blue Pill' ? Runs Forth e.g. Mecrisp you just need to add the MAX3232 and sounds like not a lot of Forth code.
Mike

That's a board. We already have a board with a MAX3232 kludged onto it. I shouldn't say "kludged" onto it, but it's on perfboard with a mating connector to an Arduino nano. This is having problems and the guy who designed it doesn't have time to debug it. I guess the first couple of units worked ok, so they shipped half a dozen. But some two or three dozen are needed to finish off the initial job. At some point this "fix" to the sensor messages could be added to the recipient, but not just now. In fact, the real problem is that the software receiving this downstream, doesn't need the headers and tosses them, but barfs if they aren't there to be tossed. I've tried to explain how simple a fix this would be, but the guy in charge is just being very cautious about this software.

If you'd like to design this and produce a box that does the translation, you can sell a few dozen. What's your price?

[[e-mail]]

Post by Lorem Ipsum
I'm looking for an MCU that can be programmed for a simple translation task, that is already mounted in a small box and ready for field use.
The task is fairly simple. The main part is to receive a once per second message of around 50 characters, so low data rate, at 9,600 bps. There is a date and a time field that needs to be modified (adding colons and slashes to make them more human readable). The modified message is transmitted on the other port, also at 9,600 bps. Every 20 messages a header of 5 lines, again around 50 char per line, is sent. Still a low data rate, relative to the ~1,000 cps that can be sent.

What does "field use" imply? Heat/Cold? Humidity? Salinity? Aggressive atmosphere?
Available power? 24/7 operation?

[Lorem Ipsum]

Post by ***@arcor.de

Post by Lorem Ipsum
I'm looking for an MCU that can be programmed for a simple translation task, that is already mounted in a small box and ready for field use.
The task is fairly simple. The main part is to receive a once per second message of around 50 characters, so low data rate, at 9,600 bps. There is a date and a time field that needs to be modified (adding colons and slashes to make them more human readable). The modified message is transmitted on the other port, also at 9,600 bps. Every 20 messages a header of 5 lines, again around 50 char per line, is sent. Still a low data rate, relative to the ~1,000 cps that can be sent.

What does "field use" imply? Heat/Cold? Humidity? Salinity? Aggressive atmosphere?
Available power? 24/7 operation?

Nothing onerous. It is in a US water treatment facility along with other commercial gear. It can't get below 0°C because the water would freeze. The tank has fish in it, so they would not like it too hot. I don't think the humidity is condensing. No acid spray or salt spray that I know of. Power is 120VAC, 60 Hz. The current unit is on a 5V wall wart. 24/7, yes. One addition that is desired is a power LED and a data received LED. The data received LED could be the equivalent of a one shot to light the LED for a quarter second whenever data comes in. Since that is once per second, it would blink. The data received is about 50 chars, so ~50 ms at 9600 bps.

[[e-mail]]

Post by Lorem Ipsum
I'm looking for an MCU that can be programmed for a simple translation task, that is already mounted in a small box and ready for field use.
This has been prototyped using an Arduino, with a MAX3232CPE chip on a perf board and a custom, 3D printed enclosure. But, about half the units exhibit a problem, where at some point, they will go haywire and garble the transmitted data.

Looks more like an electrical signal conditioning problem than defective units.
Did you hook up an oscilloscope or transient recorder? It this is too much I'd start with
RC filtering and capping peaks with (zener) diodes.

[Lorem Ipsum]

Post by ***@arcor.de

Post by Lorem Ipsum
I'm looking for an MCU that can be programmed for a simple translation task, that is already mounted in a small box and ready for field use.
This has been prototyped using an Arduino, with a MAX3232CPE chip on a perf board and a custom, 3D printed enclosure. But, about half the units exhibit a problem, where at some point, they will go haywire and garble the transmitted data.

Looks more like an electrical signal conditioning problem than defective units.

Don't know, don't care. I don't have time to mess with it. I'm happy tossing out the existing gadget. One issue I found off the bat was the lack of a range check on the character counter on the incoming data, that is used to index into the buffer array. He's also triggering the line count (and so the added header every 20 lines) by finding a \r. The input appears to contain \r\n, although I can't verify this as yet. Still, I've seen \n only, but not \r only.

Post by ***@arcor.de
Did you hook up an oscilloscope or transient recorder? It this is too much I'd start with
RC filtering and capping peaks with (zener) diodes.

Removing the translator and connecting the upstream and downstream units works with the older sensors which include the header. The sensor is a very expensive device, so not suspect. The cable is not long.

When the box stops working, it stays not working until it is reset by unplugging it.

I'm not in a position to debug the thing. The designer is too busy to debug it. If I could find a damn unit in a case with the right ports, I might take the time to write some code. As you say, it ain't much. In fact, I'm told the date and time formatting I thought was needed, is not. But we are waiting for conformation in the form of a file of data output from the new sensor.

If you'd like to do this job, let me know. Or if you know of an off the shelf MCU with at least one serial port all in a box, let me know and I'll program the thing. I'd love to use Forth, but that's not good for the customer because there are so few people who know how to use it.

[[e-mail]]

Post by ***@arcor.de

Post by Lorem Ipsum
I'm looking for an MCU that can be programmed for a simple translation task, that is already mounted in a small box and ready for field use.
This has been prototyped using an Arduino, with a MAX3232CPE chip on a perf board and a custom, 3D printed enclosure. But, about half the units exhibit a problem, where at some point, they will go haywire and garble the transmitted data.

Looks more like an electrical signal conditioning problem than defective units.

Don't know, don't care. I don't have time to mess with it. I'm happy tossing out the existing gadget. One issue I found off the bat was the lack of a range check on the character counter on the incoming data, that is used to index into the buffer array. He's also triggering the line count (and so the added header every 20 lines) by finding a \r. The input appears to contain \r\n, although I can't verify this as yet. Still, I've seen \n only, but not \r only.

Post by ***@arcor.de
Did you hook up an oscilloscope or transient recorder? It this is too much I'd start with
RC filtering and capping peaks with (zener) diodes.

Removing the translator and connecting the upstream and downstream units works with the older sensors which include the header. The sensor is a very expensive device, so not suspect. The cable is not long.
When the box stops working, it stays not working until it is reset by unplugging it.
I'm not in a position to debug the thing. The designer is too busy to debug it. If I could find a damn unit in a case with the right ports, I might take the time to write some code. As you say, it ain't much. In fact, I'm told the date and time formatting I thought was needed, is not. But we are waiting for conformation in the form of a file of data output from the new sensor.
If you'd like to do this job, let me know. Or if you know of an off the shelf MCU with at least one serial port all in a box, let me know and I'll program the thing. I'd love to use Forth, but that's not good for the customer because there are so few people who know how to use it.

Power is not the problem, nobody's got time, guys don't know and don't want Forth ...

Speaks for some board running MicroPython plus some serial to USB adapters. Choice of boards here:
https://micropython.org/download/

But since nobody's got time ....

[Lorem Ipsum]

Crap! I posted about this in several forums and got nibbles from a few, including here. I stopped replying because the waters got more murky rather than more clear. So I wanted to wait until I had something more accurate to provide people.

Meanwhile, someone sent a reply by a means I don't recall. He specifically mentioned that he had designed a product in a similar enclosure to what I want. I think he even mentioned it being IP67. But durned if I can find it. EEVBLOG has the worst messaging facility. You get to see the messages coming to you, but not the messages you send. So no conversational context!

If this jogs anyone's memory, please contact me again.

[Michael Smart]

Crap! I posted about this in several forums and got nibbles from a few, including here. I stopped replying because the waters got more murky rather than more clear. So I wanted to wait until I had something more accurate to provide people.
Meanwhile, someone sent a reply by a means I don't recall. He specifically mentioned that he had designed a product in a similar enclosure to what I want. I think he even mentioned it being IP67. But durned if I can find it. EEVBLOG has the worst messaging facility. You get to see the messages coming to you, but not the messages you send. So no conversational context!
If this jogs anyone's memory, please contact me again.

Rick

I emailed you directly to your gmail (from my gmail) you can contact me there without all the other interruptions .

Mike

[Lorem Ipsum]

Post by Michael Smart

Crap! I posted about this in several forums and got nibbles from a few, including here. I stopped replying because the waters got more murky rather than more clear. So I wanted to wait until I had something more accurate to provide people.
Meanwhile, someone sent a reply by a means I don't recall. He specifically mentioned that he had designed a product in a similar enclosure to what I want. I think he even mentioned it being IP67. But durned if I can find it. EEVBLOG has the worst messaging facility. You get to see the messages coming to you, but not the messages you send. So no conversational context!
If this jogs anyone's memory, please contact me again.

Rick
I emailed you directly to your gmail (from my gmail) you can contact me there without all the other interruptions .
Mike

Yes, I finally found your email. The email address on this account is for spam catching. People are supposed to remove the "deletethisbit" bit. lol

Anyway, I've sent a reply. Thanks.

To anyone else interested, I'm meeting with my brother again this weekend and should be able to finalize what needs to happen. Thanks for your interest.

[Paul Rubin]

Post by Lorem Ipsum
Meanwhile, someone sent a reply by a means I don't recall. He
specifically mentioned that he had designed a product in a similar
enclosure to what I want. I think he even mentioned it being IP67.

I remember posting about an RS232 switch box that looked like it was in
the right kind of enclosure. It was a metal box (not waterproof) about
the right size, looked ok, and it had three DB9 connectors and a rotary
knob. It cost around $20. So you could use the existing holes and DB9
connectors, and maybe have to drill one more hole for the power port.
I can find it again if you want, but I think the search term was "RS232
switch" or "DB9 switch" or something like that.